Can Masticatory Electromyography be Normalised to Submaximal Bite Force?

CRAWFORD, Susie <http://orcid.org/0000-0001-7871-292X>, BURDEN, A. M., YATES, J. M., ZIOUPOS, P. and WINWOOD, K.

Available from Sheffield Hallam University Research Archive (SHURA) at:

http://shura.shu.ac.uk/13797/

This document is the author deposited version. You are advised to consult the publisher's version if you wish to cite from it.

Published version

CRAWFORD, Susie, BURDEN, A. M., YATES, J. M., ZIOUPOS, P. and WINWOOD, K. (2015). Can Masticatory Electromyography be Normalised to Submaximal Bite Force? Journal of Oral Rehabilitation, 42 (5), 323-330.

Copyright and re-use policy

See http://shura.shu.ac.uk/information.html

Sheffield Hallam University Research Archivehttp://shura.shu.ac.uk

Can Masticatory Electromyography be Normalised to Submaximal Bite Force?

Susanna R. Crawford1, Adrian M. Burden

1, Julian M. Yates

2, Peter Zioupos

3, Keith Winwood

1

1 Dept. of Exercise & Sport Science, Manchester Metropolitan University, Crewe, CW1 5DU, UK 2 School of Dentistry, Manchester University, M13 9PL, UK 3 Biomechanics Laboratories, CFI, Cranfield University, Defence Academy of the UK, Shrivenham,

SN6 8LA, Corresponding author: Miss S.R. Crawford, Dept. of Exercise & Sport Science, Manchester

Metropolitan University, Crewe, CW1 5DU, UK. Email: s.crawford@mmu.ac.uk

Running Head: Normalise EMG to Submaximal Bite Force

Article category: Original research article

Abstract

The combination of bite force and facial electromyography (EMG) provides an insight into the

performance of the stomatognathic system, especially in relation to dynamic movement tasks.

Literature has extensively investigated possible methods for normalising EMG data encapsulating

many different approaches. However, bite force literature trends towards normalising EMG to a

maximal voluntary contraction (MVC), which could be difficult for aging populations or those with poor

dental health or limiting conditions such as temporomandibular disorder. The objectives of this study

were to (i) determine whether jaw-closing muscle activity is linearly correlated to incremental sub-

maximal and maximal bite force levels, and (ii) assess whether normalising maximal and submaximal

muscle activity to that produced when performing a low submaximal bite force (20N) improves

repeatability of EMG values. Thirty healthy adults (15 male, 15 female; mean age 21±1.2 years) had

bite force measurements obtained using a custom-made button-style compression load cell. Masseter

and anterior temporalis muscle activities were collected bilaterally using surface EMG sensors whilst

participants performed maximal biting, and three levels of submaximal biting. Furthermore, a small

group (n=4 females) were re-tested for reliability purposes. Coefficients of variation and intraclass

correlation coefficients showed markedly improved reliability when EMG data were normalised

compared to non-normalised. This study shows that facial EMG may be successfully normalised to a

very low bite force. This may open possibilities for comparisons between at-risk sample groups that

may otherwise find it difficult to produce maximal bite force values.

Keywords: Normalisation, Masticatory muscles, Masseter muscle, Temporal muscle, Muscle activity,

Bite force.

Introduction

Bite force and masticatory muscle activity can be used to assess the functional performance

of the jaw within research studies, and may be of some use within clinical studies. Previous

investigations have combined bite force measurements with electromyography (EMG) to explore

differences in masticatory muscle symmetry (1), masticatory function of participants with different

facial types (2, 3), and masticatory function of healthy individuals versus those with a limiting condition

such as temporomandibular disorder or migraines (4).

Clinical and research investigations using muscle activity and masticatory tasks have varied

considerably in their approach to measuring and analysing EMG data. Not only do the recording

systems, electrode parameters, and operational processes differ amongst EMG studies (5), the

techniques for processing and normalising EMG data in craniofacial and bite force research differs

depending on the purpose of each investigation. Studies investigating chewing, speech, or other

common submaximal tasks have tended to employ intercuspal Maximal Voluntary Contractions

(MVCs) as a means of normalising EMG data (2). Although this approach allows for experimental

tasks to be compared as a proportion of a maximal task, the MVC may underestimate true muscular

contraction ability due to pain and/or discomfort experienced when attempting to generate ’maximal’

bite force with bare dentition (6). Studies investigating bite force and clenching have used a device or

dampening material to produce the MVC. For example, Ferrario et al. (1) asked participants to bite

maximally on cotton rolls to normalise EMGs from intercuspal MVCs and clinical movements. In a

later study the same group (7) acknowledged that normalising bite force to an action, whether it be

maximal or submaximal, performed on a different surface to the experimental conditions, may incur a

greater level of variability. These functional differences may be due to changes in muscle length at

mouth opening height (8, 9), the level of protection of occlusal surfaces to reduce discomfort (6), or

stability and position of bite force devices (10). Using a reference voluntary contraction, such as a

nominal bite force level of 98N, to normalise EMG data from simple chewing tasks is one possible

technique (11). Employing reference voluntary contractions has become a commonly used practice in

EMG studies when participants are unable or unwilling to perform MVCs (12, 13).

Some previous studies have presented facial EMG data as absolute values rather than normalising

them to a common EMG signal (4, 7, 14). However, the amount of variation due to differences in sex,

skin thickness, and electrode placement can be reduced if the results are normalised. In the absence

of normalisation, Kemsley et al. (15) reported higher EMG inter-volunteer versus intra-volunteer, and

inter-session variation during chewing tasks. Thus, a well-controlled EMG protocol is fundamental to

increase results reproducibility (5).

This study investigated the suitability of an alternative sub-maximal bite force normalisation process.

The objectives of the study were to: (i) evaluate whether jaw-closing muscle activity is linearly

correlated with incremental sub-maximal and maximal bite force levels, and (ii) assess whether

normalising maximal and submaximal muscle activities to that produced when performing a low

submaximal bite force level (i.e., 20N) reduces between session experimental variability in a small

cohort.

Materials and Methods

Ethical approval was obtained from the Department of Exercise and Sport Science Ethics Committee,

Manchester Metropolitan University. Exclusion criteria for the study were a history of facial fracture or

surgery, current or recent orthodontic treatment, dental treatment within 6 months, musculoskeletal

disease, long-term parafunctional habits such as bruxism, temporomandibular dysfunction, or

masticatory pain. Thirty white Caucasian participants were recruited (15 males and 15 females, age

range 18-25 years, mean age 21.0 ±1.9 years). All participants gave written informed consent before

participating in the study.

Figure 1: Custom made bite force device set up.

Reliability Measures

A participant sub-cohort (n=4 females) were invited to repeat the testing session 6 months later. All

repeat participants complied with the previous inclusion criteria at the time of testing. Each participant

followed the same protocol as previously detailed with no evident learning effect.

Bite force measurements were obtained using a custom-made device consisting of two button-style

1kN compression load cells (Omega Engineering, Manchester UK) that were 3.8 mm height × 13 mm

diameter, each fitted with a 2-mm-thick stainless steel disk of the same diameter situated on top of

the button. The load cell with disk were then sandwiched between two 15 mm-diameter, 1.5-mm-thick

ethyl vinyl acetate (EVA) disks (Bracon Dental Supplies, East Sussex, UK) and held together with

electrical insulation tape. Each device (8.8 mm total height) was inserted into a latex-free vinyl sleeve,

which provided additional waterproofing for the load cell and the initial 40-50-mm of the attached wire

(Figure 1). The bite force devices were then connected to a DelSys® Bagnoli 8-channel EMG system

(Delsys Inc., Boston MA, USA) through a custom-made amplifier box. Jaw-closing muscle activity and

bite force were measured simultaneously using DelSys® EMGWorks Acquisition software (Delsys,

Inc.). Prior to testing, the bite force devices were calibrated using an LRX Plus Materials Testing

Machine 5kN (Lloyd Instruments Ltd., Hampshire, UK) under compressive loading, with a custom

made cylindrical compression loading arm of 18 mm-diameter. Each device was calibrated from 0-

800N, at incremental static compressive loads of 50N. Differential surface electrodes (Delsys, Inc.)

consisting of silver bars (10 mm × 1 mm) with an inter-electrode distance of 10 mm were used to

detect raw EMGs. The sensors including pre-amplifiers had an input impedance >1015 Ω, noise 1.2 µV

root mean square (RMS), and CMRR (Common Mode Rejection Ratio) -92 dB.

Figure 2: EMG sensor placement.

Procedures

Sensors were placed bilaterally over the main portion of the masseter, 20 mm from the inferior edge

of the mandibular angle, in a straight line to the point where the frontal bone external angular process

meets the malar bone frontal process. This line runs parallel to the superficial masseter muscle fibres.

Sensors were also placed bilaterally on the anterior portion of the temporalis, on the skin covering the

junction of the external angular process with the malar bone frontal process. The sensors were placed

posterior and parallel to the eyebrow line, so that the electrode bars were perpendicular to the anterior

temporalis muscle fibres (Figure 2). Positioning of the electrodes was verified through palpation

during clenching. Prior to attaching the electrodes, the areas of skin were shaved and cleaned with an

alcohol wipe.

During testing, participants sat upright facing a monitor positioned at eye level. EMGs were recorded

at rest, maximal- and sub-maximal bite force. First, 10-second recordings were made during complete

rest, gentle occlusion, and occlusion with the bite force devices held simultaneously on the left and

right sides, between the upper and lower molar dentition. This was used to identify anyone with high

levels of resting muscle activity, of which there were none. The participants placed the bite force

devices between their molar teeth, where they were most able to bite evenly across the two devices.

Participants then performed three repetitions of maximal voluntary biting held for 2-3 seconds, with

adequate rest between repetitions. During this phase, the investigator provided positive

encouragement whilst the participants viewed the feedback monitor, which displayed their bite force.

The investigator noted the maximum voluntary bite force (MVBF) from all repetitions and calculated

75%, 50%, and 25% force values. Subsequently, each participant then completed three sub-maximal

bite force recordings; with a horizontal target line placed on the feedback screen at 75%, 50%, and

25% MVBF in turn. Each participant was instructed to clench for ~2 seconds to reach the intended

target and then to relax for ~2 seconds repeatedly for a period of 20 seconds. This period was chosen

to allow participants enough time to identify the correct level of force and tempo to use. Finally, each

participant repeated the task to a 20N bite force target.

Data Reduction and Analysis

The DelSys EMGWorks® analysis software (Delsys Inc.) facilitated simultaneous analysis of two bite

force and four muscle activity channels. The RMS of each repetition was processed using a 0.3 s

moving window. The maximal and submaximal bite force values for both left and right sides, were

identified for each participant from a 0.15-0.20 s period. The values obtained from the clench-relax

test were selected from the first 5 peaks, to allay the effects of fatigue. These values were exported to

Microsoft Excel together with the synchronised mean EMG values for all four muscles (left and right

temporalis; left and right masseter) from the same 0.15-0.20s period.

Processed EMG data from each task were normalised by dividing these values by the muscle activity

recorded during a 20N bite force. Initially, left and right muscles were normalised separately.

Comparisons of left- and right-side data using a t-test (SPSS statistical analysis software v.19; IBM

Corp., Chicago, IL, USA) found no significant differences (p >0.05) for bite force or muscle activity in

either muscle, in both males and females. Therefore, left and right activities for each muscle were

pooled for all individuals. Mean normalised EMG data at every bite force level (25%, 50%, 75%, and

Max), were plotted against the mean non-normalised data for each muscle, and grouped according to

subject gender (Figures 3 and 4). Pearson’s correlation coefficients were calculated (SPSS statistical

software) for each muscle across the four levels of bite force, subdivided by gender and

normalisation.

Reliability Analysis

Individual and group Coefficients of Variation (CV) were calculated for non-normalised and

normalised EMG data between both testing sessions using Equation 1:

=

Equation 1: Coefficient of Variation where σ is standard deviation and µ is the mean of the sample.

The average CV was calculated separately for the left and right sides, at each maximal and

submaximal level, for both muscles. Furthermore, a two-way mixed model Intraclass Correlation

Coefficient (ICC) was calculated with SPSS statistical software, using pooled left and right data, at

each maximal and submaximal level for both muscles.

Results

Figures 3 and 4 show a linear relationship between jaw elevator muscle activity and incremental sub-

maximal and maximal bite force levels, in both male and female cohorts.

Figure 3: <25-yr-old Male cohort masseter and temporalis muscle activities (Mean ±SD) at each bite force level. Key: Mean (±SD) Masseter; Temporalis EMG normalised to 20N bite force (%);

Masseter; Temporalis non-normalised EMG values, (µV).

Despite small differences between normalised and non-normalised data, each data set had a high

Pearson’s correlation coefficient (r>0.94), indicating that the vast majority of variance in the muscle

activity is attributable to the sequential changes in bite force level.

Figure 4: <25-yr-old Female cohort muscle activities (Mean ±SD) at each bite force level. Key:

Masseter and Temporalis EMGs normalised to 20N bite force (%); Masseter and Temporalis non-normalised EMG values (µV).

Reliability Results

The between-session CVs (Table 1) predominantly demonstrate reduced variance when the EMGs

were normalised, with the only exception being the anterior temporalis muscle at 75% of maximal bite

force. The ICCs (Table 1) indicate an improved reliability between testing sessions when data are

normalised, with the exception of the anterior temporalis muscle at maximum bite force.

Coefficient of Variation (%)

Muscle Bite Force

Level Left Non-

Normalised Left

Normalised Right Non-Normalised

Right Normalised

Masseter Max 60.5 27.9 64.3 40.1

75% 43.6 31.0 49.6 40.0

50% 54.4 44.8 55.0 48.0

25% 59.1 24.0 58.9 27.1

Anterior Temporalis

Max 51.3 39.4 80.1 48.2

75% 50.2 59.7 55.5 47.4

50% 56.1 52.0 66.0 58.9

25% 55.7 32.2 74.5 53.1

Intraclass Correlation Coefficient

Muscle Bite Force

Level Non-

Normalised

Normalised

Masseter Max -0.366 0.169

75% -1.171 0.342

50% -0.271 0.245

25% -0.187 0.698

Anterior Temporalis

Max 0.265 -0.409

75% -0.92 0.672

50% -1.115 0.434

25% 0.29 0.566

Table 1: Coefficients of Variation (CV) and Intraclass Correlation Coefficients (ICC) for each muscle and bite force level.

Discussion

This study examined an alternative method for normalising EMG data recorded from the masticatory

muscles during maximal and submaximal biting tasks. We expected jaw-closing muscle activity would

have a linear relationship with incremental sub-maximal bite force levels, up to maximal bite force,

and that normalising EMG data to a low bite force (20N) would reduce the amount of variation

between testing sessions.

The results (Figure 3 & 4) show that jaw-closing muscle activity is linearly correlated with incremental

submaximal and maximal bite force levels, across male and female groups and masticatory muscles.

Pearson r-values were >0.94 for all comparisons, indicating very strong positive correlations between

muscle activity and 25%, 50%, 75% and 100% bite force intervals. These findings are an

improvement on previous work that found similar positive correlations up to 80% of maximal biting

(16) and are similar to the findings of Ferrario et al. (7) who reported a correlation >0.964 for bite force

and submaximal EMG in healthy young participants.

The present study showed that normalising to a low bite force (20N) decreased the CV and increased

the ICC in most comparisons, in a small cohort (n=4). The majority of published bite force and EMG

studies normalise to an MVC, as the focus of their investigations were submaximal tasks such as

chewing (2), clinical movements of the jaw (17), or biting at prescribed levels of force (9). The present

study used a low bite force (20N) reference value for normalisation, which allowed for successful

normalisation from 25%-100% bite force. Other researchers have used MVCs performed on

transducers or dynamometers to normalise an array of submaximal tasks on different biting (or non-

biting) surfaces (4, 17). Although biting different surfaces may create greater variation in EMG or bite

force results (1), normalisation is necessary for comparison of individual versus group results but also

enables researchers to compare their results with prior studies (18). The present study used the same

bite force devices throughout all experimental tasks, therefore potential variability caused by mouth

opening height and biting technique was reduced. Furthermore, through normalisation, these results

are suitable for cross-group comparisons. Other studies using EMG have employed normalisation to

submaximal bite force levels for analysing everyday tasks. Burnett et al. (19) found that normalising

posterior and posterolateral neck muscles to 60% MVC was reliable for both surface and

intramuscular EMG electrodes. Similarly, in non-facial EMG, Healey et al. (13) normalised paraspinal

muscle activity in participants suffering from chronic lower-back pain, to a reference voluntary

contraction obtained whilst they held a weight outstretched. . Although the reference contraction was

recorded during a different movement to the experimental conditions, it prevented any additional pain

or discomfort to the participants that may have been caused by performing a MVC. Saifuddin et al.

(11) measured EMG of the masticatory muscles during daily tasks such as chewing gum, sleeping,

and eating a meal. Similar to the present study, they normalised muscle activity to a chosen

submaximal level (a 98N bite force), which reduced mealtime EMG variation between sessions.

The current study used CVs and ICCs to quantify the reliability of the normalisation process. The CVs

markedly decreased when EMG data were normalised, with one exception highlighted in Table 1.

Moreover, the ICC results improved dramatically when EMG data were normalised. Studies that have

measured EMG data reliability, regardless of anatomical position, have employed a number of

statistical analyses to quantify the repeatability of the measurement technique: ICCs, CVs, standard

error of measurement, and repeated measures analysis of variance are commonly used (20-23).

Burdette & Gale (24) reported between-session reliability (Pearson’s r-values) for the masseter

muscle ranging between 0.56-0.65 and 0.33-0.48 for the temporalis. Although they used interclass

rather than intraclass correlation statistics for reliability measures, their results indicated increased

variability within the anterior temporalis versus the masseter muscle, which they attributed to the

temporalis’ role in maintaining mandibular postural rest. Greater variability found in this and the

present study may indicate innate differences in muscle activity between the elevator muscles.

Suvinen et al. (25) presented between-day ICC values for the masseter and anterior temporalis

muscles during mouth opening and closing tasks, with ICC results ranged from 0.877-0.899 during

clenching. The present study observed a greater ICC range and lower ICC values (0.16-0.69)

compared to Suvinen et al. (25), which could be explained by differences in EMG equipment and

placement.

In conclusion, normalising EMG data to a submaximal bite force level of 20N, highlighted a linear

relationship of jaw elevator muscle activity with sub- and –maximal bite force levels. Normalisation

successfully reduced between-session variability in comparison to non-normalised data. The

prescribed low bite force of 20N will facilitate inter-group comparisons and reduce natural variations in

masticatory muscle activity. This will prove particularly useful when studying bite force/EMG

relationships in patients with musculoskeletal conditions or in ageing populations. This study shows

that normalising EMG values to a reference level other than MVC, a technique which has been used

in other disciplines that utilise EMG (12)(13), can successfully be applied to dental and craniofacial

research with good effect.

Acknowledgements

Ethical approval was granted by the Manchester Metropolitan University ethics committee. The study

was funded by the Institute for Performance Research, MMU. The authors report no conflict of

interest. We wish to thank all participants for taking part in the study and our special thanks go to Mr

Garry Pheasey, Mr Rob Perkins and Mr Grant Rockley for their technical support.

References 1. Ferrario VF, Sforza C, Colombo A, Ciusa V. An electromyographic investigation of masticatory muscles symmetry in normo-occlusion subjects. Journal of Oral Rehabilitation. 2000;27(1):33-40. 2. Gomes SGF, Custodio W, Jufer JSM, Del Bel CAA, Garcia RCMR. Mastication, EMG activity and occlusal contact area in subjects with different facial types. Cranio: The Journal Of Craniomandibular Practice. 2010;28(4):274-9. 3. Hara A, Hara H, Uehara M, Imamura N, Ioi H, Nakata S, et al. The relationship between the craniofacial morphology and the fatigability of the masseter muscle during isometric contraction. Orthodontic Waves. 2010;69(3):85-91. 4. Burnett CA, Fartash L, Murray B, Lamey PJ. Masseter and temporalis muscle EMG levels and bite force in migraineurs. Headache. 2000;40(10):813-7. 5. Castroflorio T, Bracco P, Farina D. Surface electromyography in the assessment of jaw elevator muscles. Journal of Oral Rehabilitation. 2008;35(8):638-45. 6. Bakke M. Bite Force and Occlusion. Seminars in Orthodontics. 2006;12(2):120-6. 7. Ferrario VF, Sforza C, Zanotti G, Tartaglia GM. Maximal bite forces in healthy young adults as predicted by surface electromyography. Journal of Dentistry. 2004;32(6):451-7. 8. Paphangkorakit J, Osborn JW. Effect of jaw opening on the direction and magnitude of human incisal bite forces. Journal of Dental Research. 1997;76(1):561-7. 9. Rues S, Schindler HJ, Türp JC, Schweizerhof K, Lenz J. Motor behavior of the jaw muscles during different clenching levels. European Journal of Oral Sciences. 2008;116(3):223-8. 10. Koc D, Dogan A, Bek B. Bite force and influential factors on bite force measurements: a literature review. European journal of dentistry. 2010;4(2):223. 11. Saifuddin M, Miyamoto K, Ueda HM, Shikata N, Tanne K. A quantitative electromyographic analysis of masticatory muscle activity in usual daily life. Oral Diseases. 2001;7(2):94-100. 12. Burden A. How should we normalize electromyograms obtained from healthy participants? What we have learned from over 25years of research. Journal of Electromyography and Kinesiology. 2010;20(6):1023-35. 13. Healey E, Fowler N, Burden A, McEwan IM. The influence of different unloading positions upon stature recovery and paraspinal muscle activity. Clinical biomechanics. 2005;20(4):365-71. 14. Tecco S, Caputi S, Festa F. Electromyographic activity of masticatory, neck and trunk muscles of subjects with different skeletal facial morphology–a cross‐sectional evaluation. Journal of oral rehabilitation. 2007;34(7):478-86. 15. Kemsley EK, Defernez M, Sprunt JC, Smith AC. Electromyographic responses to prescribed mastication. Journal of Electromyography and Kinesiology. 2003;13(2):197-207. 16. Hosman H, Naeije M. Reproducibility of the normalized electromyographic recordings of the masseter muscle by using the EMG recording during maximal clenching as a standard. Journal of Oral Rehabilitation. 1979;6(1):49-54. 17. Siéssere S, de Sousa LG, Lima NA, Semprini M, de Vasconcelos PB, Watanabe PCA, et al. Electromyographic activity of masticatory muscles in women with osteoporosis. Brazilian Dental Journal. 2009;20(3):237-42. 18. Hertel J, Earl J, Tsang K, Miller S. Combining isometric knee extension exercises with hip adduction or abduction does not increase quadriceps EMG activity. British journal of sports medicine. 2004;38(2):210-3. 19. Burnett A, Green J, Netto K, Rodrigues J. Examination of EMG normalisation methods for the study of the posterior and posterolateral neck muscles in healthy controls. Journal of electromyography and kinesiology. 2007;17(5):635-41. 20. Weir JP. Quantifying test-retest reliability using the intraclass correlation coefficient and the SEM. The Journal of Strength & Conditioning Research. 2005;19(1):231-40. 21. Knutson LM, Soderberg GL, Ballantyne BT, Clarke WR. A study of various normalization procedures for within day electromyographic data. Journal of Electromyography and Kinesiology. 1994;4(1):47-59. 22. Minshull C, Gleeson NP, Eston RG, Bailey A, Rees D. Single measurement reliability and reproducibility of volitional and magnetically-evoked indices of neuromuscular performance in adults. Journal of Electromyography and Kinesiology. 2009;19(5):1013-23. 23. Worrell TW, Crisp E, LaRosa C. Electromyographic reliability and analysis of selected lower extremity muscles during lateral step-up conditions. Journal of Athletic Training. 1998;33(2):156-62. 24. Burdette B, Gale E. Reliability of surface electromyography of the masseteric and anterior temporal areas. Archives of oral biology. 1990;35(9):747-51.

25. Suvinen TI, Malmberg J, Forster C, Kemppainen P. Postural and dynamic masseter and anterior temporalis muscle EMG repeatability in serial assessments. Journal of Oral Rehabilitation. 2009;36(11):814-20.